DOI Partnership: Distinguishing Between Human and Natural Causes of Changes in Nearshore Ecosystems Using Long-Term Data from DOI Monitoring Programs

OCS Study BOEM 2019-063

DOI Partnership: Distinguishing between Human and Natural Causes of Changes in Nearshore Ecosystems Using Long-term Data from DOI Monitoring Programs

US Department of the Interior Bureau of Ocean Energy Management Pacific OCS Region

OCS Study BOEM 2019-063

DOI Partnership: Distinguishing between Human and Natural Causes of Changes in Nearshore Ecosystems Using Long-term Data from DOI Monitoring Programs

September 2019

Authors: Daniel C. Reed, Andrew Rassweiler, and Kevin D. Lafferty

Prepared under Cooperative Agreement M11AC00012 By Marine Science Institute University of California, Santa Barbara Santa Barbara, CA 93106

US Department of the Interior Bureau of Ocean Energy Management Pacific OCS Region

DISCLAIMER Study collaboration and funding were provided by the US Department of the Interior, Bureau of Ocean Energy Management (BOEM), Environmental Studies Program, Washington, DC, under Agreement Number M11AC00012. This report has been technically reviewed by BOEM, and it has been approved for publication. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the opinions or policies of the US Government, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.

REPORT AVAILABILITY To download a PDF file of this report, go to the US Department of the Interior, Bureau of Ocean Energy Management Data and Information Systems webpage (https://www.boem.gov/Environmental-Studies- EnvData/), click on the link for the Environmental Studies Program Information System (ESPIS), and search on 2019-063.

CITATION

Reed DC, Rassweiler A, Lafferty KD. 2019. DOI Partnership: Distinguishing between human and natural causes of changes in nearshore ecosystems using long-term data from DOI monitoring programs. Camarillo (CA): US Department of the Interior, Bureau of Ocean Energy Management. OCS Study BOEM 2019-063. 117 p.

ABOUT THE COVER Ronald H. McPeak, Underwater and Coastal California Photographs, 1965-1999 (Mss 292) Department of Special Collections, Davidson Library, University of California, Santa Barbara. © Regents of the University of California. Licensed for non-commercial use under a Creative Commons Attribution- NonCommercial 3.0 Unported License.

ACKNOWLEDGMENTS This research relied on data collected by the National Park Service, the United States Geological Survey, the Partnership for Interdisciplinary Studies of Coastal Oceans, and the Santa Barbara Coastal Long Term Ecological Research project. We thank these programs and all the staff and volunteers involved in collecting and assembling the data. We also think the Santa Barbara Channel - Marine Biodiversity Observing Network project for assistance with data synthesis. We thank the many scientists who participated in some aspect of this work including Donna Schroeder, Jim Estes, David Kushner, Mike Kenner, Josh Sprague, Dan Okamoto, Mark Novak, Tim Tinker, Jenn Caselle, Ally Dubel, Li Kui, Shannon Harrer, Jacob Staines, and Katie Davis.

Contents

List of Abbreviations and Acronyms ...... ii

1 Executive Summary ...... 1

2 Project Description...... 2

2.1 Introduction ...... 2

2.2 Data acquisition, documentation and publication of DOI data sets ...... 3

2.3 Integrating DOI data sets with data from different monitoring programs to facilitate analyses across a broad range of temporal, spatial and taxonomic scales ...... 4

2.4 Patterns of spatial and temporal variability in kelp forest communities and environmental and anthropogenic factors that drive them ...... 6

2.5 Integration with other BOEM-funded studies...... 7

3 References ...... 9

Appendix A: A multi-decade time series of kelp forest community structure at the California Channel Islands ...... 10

Appendix B: A multi-decade time series of kelp forest community structure at San Nicolas Island, California ...... 36

Appendix C: DOI partnership integrated data package - Metadata ...... 53

Appendix D: Detecting human impacts in a highly variable ecosystem using long-term monitoring data .. 71

Appendix E: Scale-specific drivers of kelp forest communities ...... 93

List of Abbreviations and Acronyms

ASCII American Standard Code for Information Interchange BACI Before-After-Control-Impact BOEM Bureau of Ocean Energy Management BON Biodiversity Observation Network BOT Bottom CDIP Coastal Data Information Program CAN Canopy CINP Channel Islands National Park CNMD Canopy-Midwater dbMEM distance-based Moran’s Eigenvector Maps DOI US Department of the Interior ESP Environmental Studies Program ESPIS Environmental Studies Program Information System FUL Full KFM Kelp Forest Monitoring program LTER Long Term Ecological Research program OCS Outer Continental Shelf PCNM Principal Coordinates analysis of Neighborhood Matrices PDF Probability Density Function PISCO Partnership for Interdisciplinary Studies of Coastal Oceans RDA Redundancy Analysis RDFC Random Diver Fish Counts ROMS Regional Ocean Modelling System RPC Random Point Contact SCB Southern California Bight SNI San Nicolas Island monitoring program SST Sea Surface Temperature UPC Uniform Point Contact USFWS US Fish and Wildlife Service USGS United States Geological Survey

1 Executive Summary Monitoring and predicting the potential impacts of outer continental shelf (OCS) energy production on nearshore ecosystems requires an ability to distinguish between changes caused by natural processes and those caused by human activities. The ability to distinguish such changes in turn requires long-term, spatially extensive data to describe natural patterns of temporal and spatial variation in species abundances and the environmental factors that influence them. This is particularly true for giant kelp forests, which are highly productive and diverse ecosystems in temperate regions that fluctuate greatly in space and time. These systems are highly valued for the milieu of goods and services they provide to society and there is general interest in minimizing anthropogenic activities that adversely affect them. The purpose of this project was to partner with agencies in the Department of the Interior (DOI) to document, integrate and analyze data produced from long-term kelp forest monitoring programs to improve our understanding of the causes and consequences of change in these iconic ecosystems.

The primary objectives in this collaborative partnership were fourfold: (1) Work with two Department of the Interior agencies to assimilate, document and published their long-term (30 + years) data sets pertaining to kelp forest community structure at the northern Channel Islands and San Nicolas Island; (2) Expand the spatial scope of these data sets by integrating them with other long-term kelp forest monitoring programs in the region and with appropriately temporally and spatially scaled environmental data to produce a data resource with unrivaled temporal, spatial and taxonomic scope; (3) Analyze integrated data sets across multiple spatial and temporal scales to ascertain patterns of variation in population and community dynamics and to identify key environmental and anthropogenic factors that drive them; (4) Use the fully integrated data sets to collaborate with BOEM partners and other BOEM- funded programs on issues relevant to BOEM’s mission

The value of this project to BOEM lies in its ability to assist managers in detecting and evaluating possible impacts from offshore energy activities, and in developing options to mitigate these impacts. In addition, identification of patterns in these data sets will aid in predicting potential ecosystem impacts due to climate change and advancing adaptive management, both of which are goals central to DOI stewardship responsibilities and trust resources.

One lesson of this project is that the time and expense necessary to make data visible, easy to use and easy to combine with other data should not be underestimated, but that such efforts can have very positive results on how widely that data is accessed and used. When long term ecological monitoring continues for decades, there is a danger either that too little metadata is assembled, preventing interpretation of the data, or that so much metadata is assembled that the volume of information becomes a barrier to understanding what was done. The process of publishing data sets forces metadata to be constructed in a standard format and ensures that an appropriate level of detail is maintained. When combining data sets, we note that decisions about taxonomic resolution can present obstacles to data synthesis, especially when organisms are separated into life stages or size classes. If a new monitoring project is proposed, it is worth investing time up front to ensure that the newly collected taxa groups and size classes will be compatible with existing data sets.

We show that the combined data sets we produce have considerable potential for detecting impacts in this region. Although there are statistical challenges to detecting localized impacts, Before-After-Control- Impact analyses or similar approaches can reliably detect impacts of moderate size and severity particularly if several species are similarly impacted and thus included in the analysis. However, we also reveal that the spatial and temporal dynamics of this marine system can result in extremely high false positive rates if temporal autocorrelation is not properly accounted for. Because adjusting for such

1 temporal effects requires long-term monitoring data, this emphasizes the continued value of DOI monitoring for impact analysis in the region.

Our analyses of the drivers of spatial variation within the region indicate that moderate scale variation (65 km+) is relatively predictable based on available environmental covariates. This suggests that detection of moderate or large sized impacts such as oil spills could be further improved by considering such covariates in analyses. Unfortunately, local scale impacts such as cable burial occur on a background of high small-scale natural variability. This small-scale variability was not well predicted by any of the covariates we examined, and would rarely be co-located with existing monitoring sites. These results emphasize the challenge of reliably estimating small impacts without direct collection of ecological data before and after an anthropogenic disturbance.

2 Project Description

2.1 Introduction A primary goal of environmental impact assessments is to quantify trends in species responses following a perturbation. Evaluation or predictions of the effects of anthropogenic perturbations typically are made based upon the observed or hypothesized magnitude and direction of these trends. To be useful for impact assessment analyses, however, trends observed to occur at local spatial scales (such as near an offshore energy facility) must be placed within a larger regional context. In the case of predicted trends, these changes should reflect not only the hypothesized effects due to the perturbation, but also those effects likely to occur as a consequence of unrelated, natural changes in regional environmental conditions. Consequently, studies examining long-term temporal and large-scale temporal and spatial gradients in species composition and abundance are needed to distinguish anthropogenic effects from trends arising from natural occurring phenomena. Such temporally and spatially extensive studies provide both a regional context within which observed trends can be placed, and an opportunity to explore responses of species to both natural environmental and human-induced perturbations.

Existing offshore oil and gas activities and future renewable energy projects can directly and adversely affect nearshore ecosystems. As mandated in the Outer Continental Shelf Lands Act (43 USC §1346(b)), part of BOEM’s mission is to conduct additional studies, subsequent to leasing and developing any area or region, to provide time-series and data trend information for the purpose of identifying any significant changes in the quality and productivity of the marine environment. Ocean energy development activities that affect the nearshore may include: (1) direct alteration of habitat through the installation, maintenance, and/or removal of foundations, platforms, pipelines, cables, anchors, and other structures; (2) release of contaminants into the marine environment by oil spills and discharges; (3) decreased water quality via sediment disturbance during anchoring, dredging, etc.; and (4) onshore activities that result in erosion or spillage into the nearshore environment. Thus, in order to predict the potential environmental impacts of human activities, there is a need to describe and understand the natural dynamics of coastal ecosystems. This need is particularly pressing for State natural resource managers, because, by definition, nearshore ecosystems reside in State waters.

The purpose of this project was to document, integrate and analyze data produced from long-term monitoring programs overseen by DOI and other entities to improve our understanding of the causes and consequences of change in giant kelp forests, highly productive and diverse nearshore ecosystems that occur along the Pacific coast of the US and other temperate regions of the world. The value of this project to BOEM lies in its ability to assist managers in detecting and evaluating possible impacts from offshore

2 energy activities, and in developing options to mitigate these impacts. In addition, identification of patterns in these data sets will aid in predicting potential ecosystem impacts due to climate change and advancing adaptive management, both of which are goals central to DOI stewardship responsibilities and trust resources.

Our primary objectives in this collaborative partnership were fourfold: (1) Work with two Department of the Interior agencies to assimilate, document and published their long-term (30 + years) data sets pertaining to kelp forest community structure at the northern Channel Islands and San Nicolas Island; (2) Expand the spatial scope of these data sets by integrating them with other long-term kelp forest monitoring programs in the region and with appropriately temporally and spatially scaled environmental data to produce a data resource with unrivaled temporal, spatial and taxonomic scope; (3) Analyze integrated data sets across multiple spatial and temporal scales to ascertain patterns of variation in population and community dynamics and to identify key environmental and anthropogenic factors that drive them; (4) Use the fully integrated data sets to collaborate with BOEM partners and other BOEM- funded programs on issues relevant to BOEM’s mission.

2.2 Data acquisition, documentation and publication of DOI data sets The focus of this element of the project was on working with two DOI agencies, the National Park Service (NPS) and the United States Geological Survey (USGS) to assemble, document and publish their long-term kelp forest monitoring data sets to facilitate their use by academic, governmental and privately funded scientists.

The mission of the NPS is to manage diverse park resources unimpaired for the enjoyment of future generations. To help meet this challenge, the NPS developed the Natural Resources Inventory and Monitoring Program to provide a broad-based understanding of the status and trends of park resources as a basis for making decisions and working with other agencies and the public. The NPS initiated the Kelp Forest Monitoring (KFM) program in 1981 to determine the health of kelp forests in the Channel Islands National Park, California. The KFM program seeks to establish the normal limits of variation of population parameters, to provide early warnings of abnormal conditions, and to identify possible agents of abnormal change. Permanent transects are used to monitor 70 species of algae, fish and invertebrates or groups of taxa at 33 sites within the National Park. Sixteen original sites were established between 1982 and 1986, one site was established by a commercial fisherman with funding from Santa Barbara County in 2001, and an additional 16 sites were established in 2005 as part of a fine scale study of marine protected areas that were established by the State of California in 2003.

In order to understand the ecological consequences of translocating a southern sea otter population to San Nicolas Island, the United States Fish and Wildlife Service (USFWS) in cooperation with the University of California, Santa Cruz established the San Nicolas Island monitoring (SNI) program. The program has monitored six shallow subtidal sites around the island since 1980, and a seventh site since 1986. Surveys collect information on the abundance and size of algae, macroinvertebrates and reef fishes in permanent plots. The sites have been sampled twice a year in spring and fall with few exceptions. The SNI program is managed and directed by the USGS Western Ecological Research Center stationed at the University of California, Santa Cruz.

This portion of the project entailed: (1) Working with NPS and USGS scientists to acquire and assimilate all the data from their monitoring programs; (2) Employing data quality control and assurance measures to improve the accuracy and usability of the data; (3) Developing fully documented metadata in

3 accordance with standards endorsed by national environmental data repositories; and (4) Publishing final data and metadata products in the peer-reviewed scientific literature.

The KFM and SNI data collections consist of multiple data tables that reflect different methods for sampling different types of organisms. For each collection we produced fully documented metadata for each data table and published the two data collections (KFM and SNI) as independent data papers in the journal Ecology, one of the most highly cited ecological journals in the world. These two papers (Kushner et al. 2014 and Kenner et al. 2014) are included as Appendix A and B of this report. The published products are readily accessible to scientists and the public at large and are actively being used by the science community to provide insight into the patterns and causes of change in kelp forest ecosystems within the Southern California Bight, and to provide a general framework for understanding the processes controlling the dynamics of giant kelp forest communities throughout their broader geographic distribution.

The publication of these data sets as data papers in Ecology has resulted in increased visibility and use of the data. For example, since publication (2013-2019) the KFM data has been cited in published scientific work 23 times, at a rate of 3.8 citations per year (Google Scholar search on 9/9/19). The SNI data has been cited 11 times since the data was published, a rage of 1.8 publications per year (Google Scholar search on 9/9/19). In each case, these citation rates are much higher than prior to publication. We are aware of the KFM data being cited in publications 24 times between 1982 and 2013 (a rate of 0.8 per year; See Appendix A section V.C for a list of these publications). The SNI data was cited in publications 11 times in that period (a rate of 0.4 per year; see Appendix B section V.F for a list of these publications). For both data sets, the citation rate after we published the data was almost 5 times higher than the rate in the prior years. Some of this may reflect the natural increase in value and visibility of these data sets as they increase in temporal length, but this is still an impressive increase.

2.3 Integrating DOI data sets with data from different monitoring programs to facilitate analyses across a broad range of temporal, spatial and taxonomic scales A primary objective of the DOI partnership project was to increase the utility of large data sets and the programmatic ability of BOEM to assess changes in nearshore communities and species at risk in California through the development and curation of novel data products. The first step in this endeavor was to formalize and publish data from two long-term DOI kelp forest monitoring programs as described in Section 2.2. Once this step was completed we sought to improve the spatial resolution and the regional scope of these data sets by integrating them with an additional 11 reefs monitored by the Santa Barbara Coastal Long-Term Ecological Research Program (LTER) and 25 reefs monitored by the Partnership for Interdisciplinary Studies of Coastal Oceans (PISCO). The Santa Barbara Coastal LTER (http://sbc.lternet.edu/) is an interdisciplinary research program focused on developing a predictive understanding of the processes that structure kelp forest ecosystems under conditions of changing climate and human use. It is a member of the U. S. National Science Foundation’s LTER Network charged with providing the scientific community, policy makers, and society with the knowledge and predictive understanding necessary to conserve, protect, and manage the nation’s ecosystems, their biodiversity, and the services they provide. The LTER Network was founded on the principal that long-term research is needed to help unravel the principles and processes of ecological science, which frequently involves long- lived species, legacy influences, and rare events. PISCO (http://www.piscoweb.org/) is an academic consortium led by scientists from four core campuses: Oregon State University; Stanford University’s Hopkins Marine Station; University of California, Santa Cruz; and University of California, Santa Barbara. Its mission is to conduct long-term research to advance understanding of the coastal ocean

4 within the California Current Large Marine Ecosystem and inform management and policy. Giant kelp forests and the rocky intertidal serve as the focal ecosystems for their research.

Finding a common data structure and comparable reporting metrics is essential for integrating data sets into a common framework. This is often challenging when integrating data from different sources as methods typically differ as do the data structures. Such is the case for the KFM, SNI, LTER and PISCO data as the sampling methods, frequency and timing of data collection and taxonomic resolution were individualized to meet the mission, goals and operating budgets of the respective monitoring programs. Integrating data from KFM, SNI, LTER and PISCO required us to develop a common data framework in which all data sets overlapped, and allowed users to consider any combination of data sources in an analysis. An additional problem with past efforts to join data sets is that the combined data are only valid to the point in time that the data were joined. Therefore, we developed programming scripts that make it possible to automatically update the combined data set once the source data have been updated.

A key challenge to producing data sets was to determine what biological variables have been measured in comparable ways by each program. In particular, the taxonomic resolution and the range of species counted varied widely from program to program. Our focus here was to produce data sets that are ready for analysis; thus, it is important that each species (or taxonomic group) included be counted using comparable methods by all data sources that are being synthesized. As a consequence, the taxonomic list for each integrated data set can only include species that are shared across all source data. Each data set included lacks some taxa (or taxonomic resolution) contained in the other data sources; therefore, the greater the number of sources included, the smaller the resulting usable taxa. We therefore include five versions of the biological data here (Appendix C), a core data set including all four monitoring programs, and four additional data sets - representing all three-way combinations of these sources. Although we do not include all pairwise combinations in this report, we developed computer code (Also included in appendix C) which will produce any pairwise combination except for the LTER-SNI combination (this pair is unlikely to be used in isolation because of the geographic layout of the sites, with LTER in the north, SNI in the south and PISCO and KFM in between).

Each version of the data set includes three biological data tables, a “fish” table representing counts of fish and the area sampled, a “quadswath” table representing counts of algae and invertebrates and the area sampled, and a “benthic” table containing the percent cover of macrophytes and sessile invertebrates. We also present diversity data for each of the three types of data tables that include values for richness, Shannon, and Simpson’s diversity indices calculated from the included taxa. Throughout, data are presented from 2000 onward, as both PISCO and LTER data do not start until the early 2000s.

To supplement these biological data, we developed environmental data sets that have been processed to match the spatial and temporal scales of the biological data. The environmental data sets include information about sea surface temperature, wave climate, biomass of giant kelp, and sea surface chlorophyll in the area immediately surrounding a sampling site. Environmental data are averaged over two temporal scales; the six months prior to a sampling date, and the 12 months prior to a sampling date. Environmental data sets and metadata are included in Appendix C.

In addition to supplying analysis-ready data sets, we provide the computer codes used to produce them (Appendix C). At the top of each of these scripts are user-friendly parameters that allow the user running the script to select which of the four source data sets are included in the output (using a zero to exclude the data and a 1 to include it). The fish and invertebrate processing scripts also include a parameter that allows the user to set the balance between keeping more observations and obtaining better taxonomic resolution. These codes curate the data, and draw on a large integrated data table produced by the Santa Barbara Channel Marine Biodiversity Observing Network (http://sbc.marinebon.org/).

2.4 Patterns of spatial and temporal variability in kelp forest communities and environmental and anthropogenic factors that drive them We used the integrated data set compiled from KFM, SNI, LTER and PISCO data and publically available environmental data to explore patterns of spatial and temporal variability in giant kelp forest communities and to identify environmental and anthropogenic factors that may drive dynamics. We focused our efforts on two areas. First we conducted numerical analyses using observational surveys for a broad spectrum of nearshore marine taxa to determine how nearshore species vary in their value for estimating and detecting impacts. Second we used redundancy analysis (RDA) to independently model the spatial structure of fishes, invertebrates and macroalgae to determine the spatial scales over which different environmental factors affect their species composition and abundance. We briefly summarize these activities below and present detailed descriptions of these efforts in Appendices D and E.

2.4.1 Detecting human impacts in a highly variable ecosystem using long-term monitoring data Evaluating the effects of unexpected human impacts is a major challenge for applied ecology. When ecological monitoring data are available, a Before-After-Control-Impact (BACI) analysis is often applied, which can control for natural spatial and temporal variation to better isolate an impact and help to distinguish it from other sources of variation. The virtues of BACI-type designs are well established, but the degree to which the approach may be compromised by patchy population distributions and dynamics has not been systematically explored in nearshore marine systems. Here we quantified the potential for BACI analyses of long term monitoring data to detect anthropogenic impacts with the goal of better understanding its reliability when applied to particular species (or sets of species). We started with a spatially and temporally extensive monitoring data set from reefs in southern California and simulated impacts of different spatial-scales and severities. We found the BACI approach had substantial potential to detect local reductions in the population sizes of 28 species of fish, invertebrate and macroalgae, as long as temporal autocorrelation was accounted for in the statistical model (neglecting this temporal structure led to very high false positive rates). However, the power to detect impacts varied widely from species to species and false positives exceeded target levels in some cases because of spatial synchrony in dynamics (false positives rates were lower if impact sites were scattered across the region, rather than clustered,). Impacts were most likely to be detected in species with relatively stable, widely distributed populations, but the same factors increased false positive rates. Finally we found that combing even a small number of impacted species in our analyses greatly improved the method’s success rates, but that the choice of species to be combined was crucial. These results provide guidance to long-term monitoring projects by suggesting characteristics of species that would serve as good indicators of an impact, and by setting expectations for what sizes and severities of impacts are likely to be detectable. A paper detailing this work is in preparation for publication and is included in its entirety as Appendix D.

2.4.2 Scale-specific drivers of kelp forest communities Identifying spatial scales of variation in natural communities and the processes driving them is critical for obtaining a predictive understanding of biodiversity. This is particularly true for highly productive and diverse marine communities associated with shallow subtidal reefs distributed across fragmented landscapes. Such is the case for kelp forests in southern CA and other regions of the world. Using the integrated data sets described above, we combined long-term subtidal community surveys from 86 sites with detailed environmental data to determine what structures assemblages of fishes, invertebrates and algae at multiple spatial scales. We identified the spatial scales of variation in species composition using a hierarchical analysis based on eigenfunctions, and assessed how sea surface temperature (SST), water column chlorophyll, giant kelp biomass, wave exposure and oceanographic connectivity contributed to

6 community variation at each scale. Spatial effects occurring at multiple scales explained 60% of the variation in fish assemblages and 52% of the variation in the assemblages of invertebrates and algae. Most variation occurred over broad spatial scales (> 200 km) consistent with spatial heterogeneity in SST and oceanographic connectivity, while this latter also explained community variation at medium scales (65-200 km). Small scale (1-65 km) community variation was substantial, but not linked to any of the measured drivers. Conclusions were consistent for both reef fishes and benthic invertebrates and algae, despite sharp differences in their adult mobility. Our results demonstrate the scale dependence of environmental drivers on kelp forest communities, showing that most species were strongly sorted along oceanographic conditions over various spatial scales. Such spatial effects must be integrated into models assessing the response of marine ecosystems to climate change. A peer-reviewed paper detailing this work by Lamy et al. (2018) was published in the journal Oecologia and is included in its entirety as Appendix E.

Taken together, our results emphasize the value of multisite marine monitoring projects for detecting impacts and suggest ways that better understanding of the drivers of variability in community structure might improve impact assessment. Our analysis using existing data emphasizes that even relatively simple statistical approaches such as BACI analyses can detect impacts of moderate spatial scale such as oil spills (section 2.4.1), particularly if the disturbances impact multiple species simultaneously. Natural spatial variation occurring synchronously on scales at or above 65 km can complicate such analysis, but we find that much of this variation can be predicted based on environmental factors operating at those scales such as sea surface temperatures and waves (2.4.2). This suggests the potential value of more sophisticated impact analyses that incorporate physical covariates to account for naturally-driven changes in community structure. However, we also document high levels of small-scale community variation which were not reliably predicted by any of the drivers we examined. This small-scale variation presents a real challenge for estimating the effects of more localized disturbances such as cable burial. Our BACI analyses indicate that such small impacts would rarely be resolved with existing data because they would be unlikely to occur near one of the monitoring sites, and the high degree of local variability we observe means that we have limited ability to predict the community that would have occurred at a sites in the absence of human activity. If the location of small-scale impacts can be predicted in advance, for example with cable as with cable burial, then collecting pre-impact data on ecological communities at those locations may be the only way to obtain reliable estimates of those impacts.

2.5 Integration with other BOEM-funded studies

2.5.1 Estimates of wave energy and its effects A key integrative activity has been to combine the biological and environmental data synthesized here with wave energy estimates. The development and validation of wave energy estimates was done under a separate BOEM award, Predicting the Consequences of Wave Energy Absorption from Marine Renewable Energy Facilities on Nearshore Ecosystems (PC-13-05). This project evaluated the spatial and temporal variation in wave height for the study system in the broader context of the Southern California Bight. This helped identify two wave seasons as well as the distribution of exposed and protected sites in the region. We deployed, retrieved and processed data from a new, low-cost pressure sensor at 32 sites around the Channel Islands where long-term kelp forest monitoring occurs (a combination of KFM and SNI sites). We then compared hourly data from the sensors with hourly hindcasts from the Coastal Data Information Program (CDIP; https://cdip.ucsd.edu/). This made it possible to generate simple statistical corrections to CDIP hindcasts to help improve their accuracy. With this correction, we generated hourly hind-casts for these 32 sites back to 2000.

These high frequency wave data were then summarized at an annual timescale to better match the timing of biological sampling. This allowed us to extrapolate the results of Lafferty and Morton (2018) to product hindcasts for an additional 56 sites monitored by PISCO and LTER, bringing our total number of sites to 88. At each of these sites we produced hindcasted estimates of mean wave height, energy, and orbital velocity calculated at 6- and 12-month intervals before biotic sampling. The hind-cast wave data are publicly available on the USGS ScienceBase-Catalog (Lafferty et al. 2018) and can be used to address questions about how waves affect nearshore communities in the Southern California Bight.

Lastly, we integrated the wave hindcast data with the biological and environmental data sets described in Section 2.3 to explore the degree to which wave height and other abiotic factors drive species dynamics at the 88 sites. This integrated data set was the primary data source for our investigations on scale-specific drivers of kelp forest communities (Lamy et al. 2018; see Section 2.4.2).

2.5.2 Marine Biodiversity Observation Network

We collaborated extensively with the study A Demonstration Marine Biodiversity Observation Network (BON) for Ecosystem Monitoring (PC-15-05) in their efforts to integrate data from diverse sources across the region into a common data structure. Our work in preparing the KFM and SNI data sets for publication (Section 2.2) paved the path for the BON to produce several integrated data sets of nearshore biological data from the region (Miller et al. 2016a, b, Miller et al. 2017a, b). We also collaborated with the BON to add the wave data from Lafferty et al. (2018) to the integrated biological data set in this report and into a separate BOEM-funded integrated data table produced by the Santa Barbara Channel Marine BON.

A limitation of these BON-generated data sets for researchers seeking to use them is that the data are kept in a form as close as possible to the original, which creates substantial hurdles to using data from multiple sources. Data users must either limit themselves to a single data source within the integrated data set, or understand the details of how the data collection methods differ and make their own decisions about how to combine the data. In cases where the data sources use different methods/rules to aggregate species, life stages or size classes into a taxonomic group, the user must consider these differences and decide whether the data are comparable and if so how the data set should be aggregated. In many cases a species may be sampled by multiple methods within a single data set and the user must decide which method is more reliable and comparable with data from the other sources being combined. Scaled across hundreds of species and four different sampling programs, each with its own complex history, this decision-making process represents months of work.

As described in Section 2.3, we created curated data sets and associated data processing scripts to assist the user in overcoming the above hurdles associated with using data from multiple sources, which enables them to address their questions more immediately. This is a particularly important product for BOEM and other DOI agencies, as a primary goal of this project was to ensure that relevant nearshore biological data are ready for immediate analysis to support decision-making or impact assessment related to offshore energy production.

3 References

Lafferty KD, Morton DN. 2018. Site table and bias corrections for Coastal Data Information Program (CDIP) hind casts at the California Channel Islands: U.S. Geological Survey data release, https://doi.org/10.5066/P9FV1X7H.

Lafferty KD, Morton DN, Gotschalk CC, Henderikx F, Rassweiler A, Washburn L. 2018. Hourly wave height and period hindcasts at 32 sites throughout the Channel Islands National Park and San Nicolas Island from 2000-2017: U.S. Geological Survey data release, https://doi.org/10.5066/P91KH2Q7.

Miller R, Rassweiler A, Reed D, Lafferty K, Kui L, O'Brien M. 2016. Santa Barbara Channel Marine BON: Integrated quad and swath cover. Environmental Data Initiative. https://doi.org/10.6073/pasta/51d2db26e90d4b8687db81fb40bc58c4. data set accessed 8/19/2019.

Miller R, Rassweiler A, Reed D, Lafferty K, Kui l, O'Brien M. 2016. Santa Barbara Channel Marine BON: Integrated benthic cover. Environmental Data Initiative. https://doi.org/10.6073/pasta/bf143fa962e1edb822847bc0ee90c2f7. data set accessed 8/19/2019.

Miller R, Rassweiler A, Reed D, Lafferty K, Kui L, O'Brien M. 2017. Santa Barbara Channel Marine BON: Integrated fish. Environmental Data Initiative. https://doi.org/10.6073/pasta/1345f0148e6dfe4df9065e223b4dd783. data set accessed 8/19/2019.

Miller R, Rassweiler A, Reed D, Lafferty K, Kui L, O'Brien M. 2017. Santa Barbara Channel Marine BON: Integrated taxa. Environmental Data Initiative. https://doi.org/10.6073/pasta/d09d4bfd54e6d4e490b4cc34731d808e. data set accessed 8/19/2019.

Appendix A

All files and folders referenced in Appendix A are found in the file: OCS Study BOEM 2019-063 Appendix A files.zip

A multi-decade time series of kelp forest community structure at the California Channel Islands

David J. Kushner, Andrew Rassweiler, John P. McLaughlin, Kevin D. Lafferty

2013. Ecology 94: 2655

INTRODUCTION

The ocean waters surrounding the California Channel Islands include many shallow nearshore reefs. Giant kelp (Macrocystis pyrifera) is commonly, but not always, a dominant fixture here (Engle 1994, Cavanaugh et al. 2011). Biotic communities on these reefs vary in space due to differences in depth, substrate, temperature, nutrients, and exposure to waves. In addition to seasonal variation (not measured in this study), rocky reef communities may vary from year to year due to recruitment events, physical disturbance from waves, fluctuations in the climate (including El Ninos and shifts in the Pacific Decadal Oscillation) and community dynamics such as disease outbreaks or changes in grazer behavior (Lafferty 2004).

Humans also affect this ecosystem in space and time. Most of the area is open to commercial and recreational fishing (e.g, Schroeter et al. 2001), and it also supports other recreational activities such as diving and boating. Marine communities may also be affected by chronic and acute pollution from mainland waste disposal and adjacent offshore petroleum development. The five northern California Channel Islands are primarily under the management and jurisdiction of the National Park Service and Nature Conservancy, while the California Department of Fish and Wildlife has jurisdiction over the living marine resources within 3 miles of the island.

To assist in protecting and managing these nearshore ecosystems, the National Park Service created the kelp forest monitoring program (KFM) in 1982. Every summer since then, the KFM has tracked the populations of key species of algae, invertebrates and fishes at locations throughout the park. The resulting data on subtidal community structure and dynamics are nearly unparalleled in their completeness and spatio-temporal extent. These data have contributed substantially to management actions, such as the State of California’s decision to close several abalone fisheries in 1997 and the decision to create marine protected areas within the National Park in 2003. The data have also been the basis of a number of general ecological insights (e.g., Reed et al. 2000, Lafferty 2004, Byrnes et al. 2006).

Despite its proven usefulness, several characteristics of this monitoring data set have made it challenging to analyze. The KFM has refined its methods several times (e.g., Davis et al. 1996). These improvements have focused on obtaining the best data possible given a fixed amount of effort, but have come at some

cost to the overall consistency and simplicity of the data set. Researchers wishing to use these data have needed to work closely with the program’s extensive documentation, understanding how and when sampling size changed, what species have been counted with which methods in which years, and how choices about which species to count and at what taxonomic resolution have changed over time.

In this data paper, we have summarized observations from five different sampling methods into three core data sets that record the abundance of more than 60 key taxa over a 30-year period. We have corrected for changes in sampling method, changes in taxonomic resolution and other inconsistencies, seeking to make the data easy to use and help researchers avoid potential pitfalls. In addition to these core data sets, we present seven supplementary data sets that record the size structure of key species and provide additional information about fish and kelp densities, patterns of invertebrate recruitment and historical water temperatures. These supplementary data provide valuable information about the processes that underlie the community dynamics described in the core data sets.

The data might be used to describe how these rocky reef communities have changed over time, to evaluate local management practices and human impacts, and to detect the effects of larger-scale anthropogenic effects such as global climate change and ocean acidification. They also give considerable insight into the ecology and population status of the particular species being monitored. Regardless of the use, we strongly recommend that researchers contact the coordinator of the KFM when planning analyses to gain a better understanding of the data set and the system.

METADATA Metadata table of contents Class I. Data set descriptors A. Data set identity B. Data set identification code C. Data set description 1. Data set originator 2. Abstract D. Key words

Class II. Research origin descriptors A. Project description 1. Objectives 2. System description 3. Guidelines for use of data B. Monitoring design 1. Site selection and layout 2. Human impacts 3. Monitoring method selection 4. Choice of species to be monitored 5. Monitoring design metadata C. Data processing and presentation 1. Core community structure data sets 2. Supplementary data sets 3. Additional documentation and data

D. Monitoring methods and data 1. Core community structure data sets a. Benthic density data i. Benthic density summary procedures ii. 1 m2 quadrat monitoring methods iii. 5 m2 quadrat monitoring methods iv. Band transect monitoring methods b. Benthic cover data i. Benthic cover summary procedures ii. Random point contact monitoring methods c. Fish density data i. Fish density summary procedures ii. Visual fish transect monitoring methods 2. Supplementary data sets a. Roving diver fish count data i. Roving diver fish count methods b. Fish size-frequency data i. Fish size-frequency methods c. Invertebrate size-frequency data i. Invertebrate size-frequency methods d. Giant kelp size-frequency data i. Giant kelp size-frequency methods e. Giant kelp supplementary density data i. Giant kelp supplementary density methods f. Artificial recruitment module data i. Artificial recruitment module methods g. Subtidal temperature data i. Subtidal temperature methods

Class III. Data set status and accessibility A. Latest data update B. Latest metadata update C. Data verification D. Copyright or proprietary restrictions

Class IV. Data set structural descriptors A. Data files B. Metadata tables

Class V. Supplemental descriptors A. Location of completed data forms B. Data entry verification procedures C. Publications using the data set

Class I. Data set descriptors

A. Data set identity B. Data set identification code C. Data set description 1. Data set originator 2. Abstract D. Key words

I.A. Data set identity: Subtidal community data from the Channel Islands National Park’s kelp forest monitoring program.

I.B. Data set identification code: KushnerEtAl2013-KFMP.

I.C. Data set description

I.C.1. Data set originator Individual: David J. Kushner Role: Program Manager, Corresponding Author Organization: Channel Islands National Park Position: Marine Biologist Address: 1901 Spinnaker Dr. Ventura, CA 93001 Phone: (805) 658-5773 Email: [email protected]

I.C.2. Abstract:

Studies of temperate rocky reef communities have added much to our understanding of ecology. However, data on these reef communities can be difficult to obtain; wind, waves and poor underwater visibility often prevent research diving, and even under the best conditions only a few hours a day can be spent underwater collecting data. Here, we present data on temperate subtidal reef communities at 33 sites, almost half of which the National Park Service has sampled annually since 1982. We present core data sets describing the population dynamics of 16 fish, 37 invertebrate and 15 algal taxa. We include supplementary data sets documenting the size structure of key species, the relative abundance of all fish, the recruitment of selected invertebrates and the subtidal water temperature at each site through time. Taken together, these data provide one of the most comprehensive descriptions of nearshore reef community dynamics ever assembled.

I.D. Key words: Kelp Forest, Long-term Monitoring, California Channel Islands, subtidal reef ecology, marine protected areas

Class II. Research origin descriptors

A. Project description 1. Objectives 2. System description 3. Guidelines for use of data

II.A.1. Objectives: When the United States Congress passed legislation creating the Channel Islands National Park (CINP), the National Park Service (NPS) was charged with conducting “an inventory of all terrestrial and marine species, indicating their population dynamics, and probable trends as to future numbers and welfare” (16 USC § 410FF-2, 1980). Shortly after the park was created in 1980, the NPS initiated several ecological inventorying and monitoring programs, including the kelp forest monitoring program (KFM), focusing on the ecological communities found on shallow rocky reefs around the five northernmost Channel Islands. The KFM takes a population dynamics approach to describing the status of communities on these reefs, monitoring the abundance of selected species in the same locations every year. In addition to these records of population status the KFM collects supplementary data from each of these locations to help describe the biological and physical processes that govern community dynamics on these reefs.

II.A.2. System description: This study focuses on shallow (< 20 m deep) rocky reefs in the CINP. The waters of CINP give refuge to an ecologically diverse collection of species. The park is located at the boundary between two major biogeographical provinces: the Oregonian province to the north and the Californian to the south. The western park islands, San Miguel and Santa Rosa, are regularly subject to the cooler waters of the California Current. Waters around the eastern park islands of Anacapa and Santa Barbara come from the south along the mainland coast and are typically warmer. Around Santa Cruz Island, there is a broad transition zone influenced by waters from the north and south (Figure 1).

Prevailing winds and the bathymetry of adjacent basins also greatly influence marine communities in the park. Strong north winds and winter swells buffet the north sides of the islands, while the biota of the southern coasts reflects their more sheltered location. Upwelled nutrients from 2,000 meter-deep basins to the south and west of the park contribute to exceptionally productive food webs and create temperature regimes that differ significantly from the relatively shallow northern sides of the islands.

II.A.3. Guidelines for use of data: This data set is part of the ongoing Kelp Forest Monitoring Program. Researchers using these data are encouraged to contact the corresponding author (David Kushner) concerning questions, improvements, and, when appropriate, collaboration. The first author has been managing the National Park Service’s kelp forest monitoring program since 1992 and can provide more details on the sampling methodology and the natural history of the system. Additionally, he keeps track of past and ongoing analyses to help reduce duplication of effort (see also section V.C. Publications using the data set).

B. Monitoring design 1. Site selection and layout 2. Human impacts 3. Monitoring method selection 4. Choice of species to be monitored 5. Monitoring design metadata

II.B.1. Site selection and layout: The Channels Islands National Park’s Kelp Forest Monitoring Program began in 1982 and has grown over time to include 33 sites. A single permanent transect at each site is sampled annually, between May and October. A map of the park and monitoring site locations is given in Figure 1.

Figure 1: Map of kelp forest monitoring site locations

The initial 16 monitoring sites were selected from throughout the CINP; 13 sites were established in 1982, one in 1983 and two in 1986. As described below, site selection was not random, and, as a result, average densities for the sites should not be used to extrapolate to average conditions in the system. They are better suited for tracking changes over time, or matching biological communities to site characteristics.

The initial sites were distributed on both the north and south sides of each of the islands, and chosen to cover the east-west transition from the Californian to the Oregonian biogeographic provinces. Within these broad parameters, sites were selected based on the presence of continuous patches of rocky habitat (typically more than 100 m in long-shore extent) and because they were known to be good habitat for

giant kelp forests (i.e., giant kelp was known to be present at all sites in 1982 or shortly prior). Of the initial 16 sites, two were placed within a no-take marine reserve on Anacapa Island (#12 and #13), and a third placed in a nearby seasonal closure on Anacapa Island (#11).

In 2001, the Miracle Mile site (#21) at San Miguel Island was added to the KFM for the purpose of monitoring abalone populations. This site was selected by a commercial fisherman based on its exceptionally high density of the red abalone, Haliotis rufescens. In 2005, the Park received funding to expand its monitoring within and adjacent to several newly established marine protected areas (MPAs) at the Channel Islands. The KFM added 16 new sites located inside or adjacent to four of the newly established MPAs: Santa Barbara Island State Marine Reserve, Anacapa Island State Marine Reserve, Scorpion State Marine Reserve, and South Point State Marine Reserve. These four MPAs were chosen for additional monitoring because they were accessible, contained existing KFM monitoring sites and were in areas most likely to be impacted by fishing. New sites were established to complement existing sites so that at least three sites were inside and three adjacent to each of the four MPAs. Permanent sampling locations at these sites were placed in areas of continuous rocky reef. All 33 kelp forest monitoring sites and their locations (WGS 84) are described in Table 1. This information is also available in a format that can be read by geospatial software such as Google Earth. Links to an underwater video transect done at each site in 2010 are provided in Table 1, videos from other years are available from the NPS.

Monitoring at each site is focused on a permanent 100-meter transect, typically laid out parallel to the shoreline along a rough depth isobath. Permanent transects were used to reduce within-site variability in the time series and thus provide measurements of population dynamics with time as the primary independent variable. Each transect is marked by a 12 mm diameter lead-filled woven nylon line permanently affixed to the bedrock with stainless steel eyebolts positioned every 10 meters along the transect. Transects are relocated using GPS and diver search; while individual eyebolts have been lost and needed to be replace from time to time, the permanent line has allowed the transect location to be accurately relocated in all such cases. The transect depths at the original 16 sites vary from site to site, spanning a range from 5-18 meters. Transects at sites established in later years were placed between 7-17 meters in depth.

II.B.2. Human impacts: Many species monitored here are harvested commercially or recreationally. In particular, large and valuable fisheries exist for spiny lobster (Panulirus interruptus), California sheephead (Semicossyphus pulcher), red sea urchins (Strongylocentrotus franciscanus) and sea cucumbers (Parastichopus parvimensis). Although most of the sites are open to fishing and have no special restrictions on harvest beyond those common throughout southern California, several sites are currently within marine protected areas where harvest is restricted (Airamé et al. 2003, Davis 2005). All of the protected areas were established in 2003 except the area at the east end of Anacapa Island which predates the monitoring program and includes two of the original monitoring sites. A list of the special restrictions on harvest at each site (if any) and the dates those restrictions were implemented is given in Table 2.

II.B.3. Monitoring method selection: The great diversity of organisms and physical habitats associated with kelp forests require multiple sampling approaches to effectively document species population dynamics. Sampling methods were designed to maximize accuracy and precision while preserving an observer’s ability to efficiently sample several target species at once. Additionally, sampling techniques were designed to minimize variation among observers and impacts on the populations of organisms being monitored. Some details of the sampling techniques have been revised several times since the project started. Most notably, several methods have changed their sampling effort one or more times as the result of ongoing power analyses and review of the data. These changes in sampling intensity have been noted in the metadata and in the data sets.

II.B.4. Choice of species to be monitored: The primary objective in selecting taxa for monitoring was to provide a representative cross section of the ecological guilds present in the CINP kelp forests. There are currently 15 algal, 37 invertebrate, and 16 fish taxa specifically targeted in the monitoring, and all fish are monitored with supplementary methods. The species monitored and the data sets in which they appear can be found in Table 3. Though many common organisms are identified to the species level, some are identified only to a coarse taxonomic level (e.g. green algae).

II.B.5. Monitoring design metadata: The tables describing general monitoring methods described above are also available below as comma-separated values files.

Table 1: Monitoring sites and their coordinates

Table 2: History of special fishing restrictions

Table 3: Species monitored by method

C. Data processing and presentation 1. Core community structure data sets 2. Supplementary data sets 3. Additional documentation and data II.C.1. Core community structure data sets: We present data on kelp forest community structure in a format that can be easily analyzed, and so have taken steps to construct data sets that are consistent through time. To achieve this we have combined data from different sampling methods where possible, removed species and sites that were sporadically measured and corrected for changes in sampling effort over time. This has resulted in three summarized “core community structure data sets”: benthic density (calculated from counts of organisms in 1 m2 quadrats, 5 m2 quadrats and band transects), benthic cover (calculated from random point contact data), and fish density (calculated from visual fish transects). In each of these data sets, the abundance of each species or taxa sampled is expressed as a mean for each site on each sampling date with associated variance where appropriate. Full details on both the methods used to collect these data and the steps taken to process and summarize the raw measurements are given below. Note that the sampling design has been consistent since 1996, so users concerned about changes in sampling effort can limit their analysis to data collected after that year.

II.C.2. Supplementary data sets: In addition to the core data discussed above, the KFM collects several other forms of data that are not simple measures of abundance. We provide these data in seven separate data sets: roving diver fish counts, fish size-frequency, invertebrate size-frequency, giant kelp size- frequency, giant kelp supplemental density, artificial recruitment module data and subtidal temperature data. These data have not been summarized to a single number for each site and year, so will likely require some processing by users. Although the data are of high quality, many of these data sets are complex and we caution users of these data to carefully read the methods to understand how sampling was done.

II.C.3. Additional documentation and data: Considerable additional documentation is available for the KFM, including the sampling protocols used to guide field teams, annual reports describing changes in the system, and video transects from 1983-2011 for each site. These are available from the National Park Service website (www.nps.gov/chis). Additional data including the unprocessed field data can be obtained by contacting David Kushner, or the current superintendent of the Channel Islands National Park.

D. Monitoring methods and data 1. Core community structure data sets: a. Benthic density data i. Benthic density summary procedures ii. 1 m2 quadrat monitoring methods iii. 5 m2 quadrat monitoring methods iv. Band transect monitoring methods b. Benthic cover data i. Benthic cover summary procedures ii. Random point contact monitoring methods c. Fish density data i. Fish density summary procedures ii. Visual fish transect monitoring methods

II.D.1. Core community structure data sets

II.D.1.a. Benthic density data

Table 4A: Definition of column headers in the benthic density data

Table 4B: Definition of variables in the benthic density data

Benthic density data: ASCII file in comma separated values format

II.D.1.a.i. Benthic density summary procedures: Benthic density data are collected by counting the number of individuals in a fixed area and dividing that count by the area sampled. This form of monitoring is used to estimate the abundance of larger benthic organisms such as sea stars and kelps as well as some small benthic fish. In the KFM, three separate methods are used at each site to estimate benthic densities (1 m2 quadrats, 5 m2 quadrats and band transects, sampling 1-2 m2, 5 m2 and 40-60 m2 respectively). Because the sampling intensities for these methods have varied through time, data are summarized as the mean density (per m2) of each taxa at the site, averaged over all replicates within the site for a given year. Also included are the standard errors in density calculated across the sampling replicates, the number of replicates and the area sampled in each replicate.

Different methods are used to count different species, and in some cases the method used to count a particular species changed over time. By combining the densities estimated from each method into a single data set, we assemble continuous time series of the abundance of each species for which density estimates were made. In the few cases where the same species was sampled by multiple methods at the same site in the same year, we used the method that sampled the largest area.

Some species have been added to the monitoring protocols during the 30+ years of monitoring. Thus the absence of these species from the data early in monitoring cannot be taken as evidence of absence. For this reason, instead of a 0 or blank, the code “NA” is entered into the data set as the density for species in years they were not counted. Different size classes of several kelps were counted separately for the later portion of the monitoring. However we have summed across age classes to obtain consistent categories throughout this data set.

II.D.1.a.ii. 1 m2 quadrat monitoring methods: 1 m2 quadrats have been sampled since 1982 by counting the number of individuals of each target species occurring within a set of small quadrats at each site. Currently 12 quadrats are sampled, with pairs of quadrats placed every 8.33 m (rounded to the nearest

whole meter) along the main transect (Figure 2). Over the course of the program, sample size has varied from 12 to 40 quadrats, and the area of each replicate quadrat has varied from 1 m2 to 2 m2 (Table 5).

Table 5: 1 m2 quadrat sampling intensity by year.

Year # Quadrats Size 2 1982 30 1 m 2 1983-1984 40 1 m 2 1985-1995 20 2 m 2 1996-2011 12 2 m

Placement of quadrats is based on a stratified random design, with new points on the transect chosen every year, although the exact method of stratification has changed over time. For this monitoring method, observers place two 1 m2 frames adjacent to the main transect, so that they form a 2 m2 area (e.g., A1 and B1 in Figure 2) using the transect line as the fourth side of each quadrat. When sampling, divers place the frames, then withdraw to allow island kelp fish, gobies and other easily disturbed species to emerge from crevices before approaching the quadrats slowly. Fishes within the quadrat are counted first; additional fish that enter the quadrat after this first count are ignored. Divers search under ledges and in cracks for all target organisms but do not conduct invasive sampling (e.g., turning over rocks, sifting through sand, etc.).

Figure 2. Placement of 1 m2 quadrats under current design

II.D.1.a.iii. 5 m2 quadrat monitoring methods: 5 m2 quadrats have been sampled since 1996 by counting the number of individuals of each target species occurring within a set of 5 m2 sampling areas at each site. This sampling method was added to improve sampling of key species that are too rare to be regularly encountered in 1 m2 quadrats, but too numerous to quickly count in the larger band transects. The 100 m transect is divided into 20 segments, each associated with two 5 m by 1 m sampling areas, placed one on either side of the transect (e.g., Figure 3). Divers search the habitat thoroughly for target species, including cracks and crevices but do not conduct any invasive sampling (e.g., turning over rocks, sifting through sand, etc.).

Figure 3. Placement of 5 m quadrats along the transect line.

II.D.1.a.iv. Band transect monitoring methods: Band transects have been sampled since 1983 by counting the number of individuals of each target species occurring within a set of band transects at each site. Currently 12 bands are sampled, each covering 60 m2, but sampling intensity has varied over time (Table 6).

Table 6: Band transects sampling intensity by year.

Year # Band Transects Band Transect Size

2 1983-1984 10 40 m 2 1985-2011 12 60 m

Placement of bands is based on a stratified random design, with different points on the main transect chosen every year for sampling. At each sampling point, divers run measuring tapes 10m in each direction, perpendicular to the 100 m transect line, sampling organisms first within 1.5m of one side of the tape and then within that distance of the other tape (Figure 4, before 1985 only 1m on each side was sampled). Divers search the habitat thoroughly, including cracks and crevices, however, they do not conduct any invasive sampling (e.g., turning over rocks, sifting through sand, etc.).

Figure 4. Example of band transect placement under current design.

II.D.1.b. Benthic cover data

Table 7A: Definition of column headers in benthic cover data

Table 7B: Definition of variables in the benthic cover data

Benthic cover data: ASCII file in comma separated values format

II.D.1.b.i. Benthic cover summary procedures: Percent cover data for sessile benthic biota (e.g., bryozoans, understory macroalgae) are collected using a random point contact (RPC) method, in which a set of randomly chosen points are superimposed over the bottom and the species contacted by each point are recorded. The number of times each species was counted is divided by the total number of points to give the percent of the bottom occupied by that species. The physical substrate is also categorized during the RPC process and the cover of each substrate is recorded in this data set. The benthic cover data set contains the mean percent cover of each target organism, averaged across all RPC units sampled at that site, the standard error in percent cover calculated across the RPC units and information about the number of RPC units sampled and the number of points per unit. The sampling intensity for this method has changed through time, but by summarizing the data as an average percent cover for the site, we have corrected for this variation in effort.

Some species have been added to this method’s target species list during the 30 years of monitoring. Thus the absence of these species from the data early in the monitoring cannot be taken as evidence of absence. The code “NA” is entered into the data set as the percent cover of species in years they were not counted. Several species which were not distinguished individually early in the monitoring were later sampled

separately. However, we have aggregated such species groups in the benthic cover data so that taxonomic resolution is consistent throughout the data set.

II.D.1.b.ii. Random point contact monitoring methods: Benthic cover data are collected in RPC units of 40 points distributed around a given location on the main transect. Currently 15 units placed approximately every 6.6 m (rounded to the nearest meter) along the transect are sampled at each site, but the intensity of sampling has varied over time (Table 8). Placement of RPC units follows a stratified random design, with new locations on the transect chosen every year, although the exact method of stratification has changed over time. Again, this introduces an element of year to year variation not due to actual changes over time.

Table 8: RPC sampling intensity by year.

Year Replicates Points per replicate 1982 25 20 1983 40 10

1984 10 50 1985-1995 25 40 1996-2011 15 40

To perform this sampling, a diver uses an “RPC bar” a 1.5 m rod with 2 strings attached, each string with 5 knots. One string is 1.8 m long and the other measures 1.2 m. The long string attaches to the ends of the bar and the short string attaches 25 cm from each end. At each sampling location, the diver places the RPC bar perpendicular to the transect line (Figure 5), holds the RPC bar in place and stretches the string taut at each knot perpendicular to the bar. This forms a triangle with the bar as one edge and the knot as the opposite vertex. The diver then imagines a line running vertically through the knot from the substrate up to one meter above the substratum and identifies all organisms that intersect this imaginary line. The diver identifies all organisms under or above the knots on both strings on one side of the bar, giving a total of ten points. Next, the diver moves the strings to the opposite side of the RPC bar for another 10 points, then places the bar on the other side of the transect line and repeats the process (Figure 5).

Figure 5: Orientation of RPC bar and strings during sampling.

The diver counts each species only once per point even if multiple individuals of the same species intersect that point. Because of this method of counting multiple layers, total cover of all species can exceed 100% (and usually does), but the cover of any individual species cannot be greater than 100%. The one exception is the category “Macrocystis, Pterygophora, and Eisenia combined” (species code 2008). Each species in this category is counted separately and then these counts are added, so percent covers of as high as 300% are possible for this combined-species category (if all points intersect all three species). Attached or sessile animals providing cover are recorded as “Miscellaneous Invertebrates.” Motile invertebrates (except for Ophiothrix spiculata and Pachythyone rubra, which are counted with “Miscellaneous Invertebrates”), are not counted but are moved to determine what is underneath. The diver also identifies the substrate type encountered at each point.

II.D.1.c. Fish density data

Table 9A: Definition of column headers in fish density data

Table 9B: Definition of variables in the fish density data

Fish density data: ASCII file in comma separated values format

II.D.1.c.i. Fish density summary procedures: Fish density data are collected by counting the number of individuals in a single fixed plot at each site. This form of monitoring is used to estimate the abundance of 13 species of fish and is based on visual fish transect data. For each fish species counted in a given year, the fish density data set contains the total count of that species; because these data are based on a single transect, no variability is reported. Occasionally the same plot was sampled on multiple dates within a single year; these counts are reported as separate entries in the fish density data set. Sampling for Semicossyphus pulcher juveniles began in 1996. Thus their absence from the data early in the monitoring cannot be taken as evidence of absence. The code “NA” is entered into the data set as the density in years they were not counted.

II.D.1.c.ii. Visual fish transect monitoring methods: To count fish, a diver swims along the main 100 m transect line, counting and recording all the individuals of a core set of species within a 3 m tall x 2 m wide x 100 m long area in front of the diver. The diver records any fish seen within this area (Figure 6). Divers typically swim the 100 m transect in 5 minutes. Counts are differentiated by species and by age class—adults or juveniles, with juveniles defined as less than 10 cm in length, except for Halichoeres semicinctus and Semicossyphus pulcher for which juveniles must be less than 10 cm in length and have a white stripe, and Hypsypops rubicundus for which juveniles are not identified by size, but by the presence of blue spots. A few species are also differentiated by sex. Fish transects are conducted once or twice per summer at each site, as time allows, with a minimum of two weeks between sampling events.

Figure 6: Area covered by visual fish transects.

II.D.2. Supplementary data sets

a. Roving diver fish count data i. Roving diver fish count methods b. Fish size-frequency data i. Fish size-frequency methods c. Invertebrate size-frequency data i. Invertebrate size-frequency methods d. Giant kelp size-frequency data i. Giant kelp size-frequency methods e. Giant kelp supplementary density data i. Giant kelp supplementary density methods f. Artificial recruitment module data i. Artificial recruitment module methods g. Subtidal temperature data i. Subtidal temperature methods

II.D.2.a. Roving diver fish count data:

Table 10A: Definition of column headers in RDFC data

Table 10B: Definition of variables in the RDFC data

RDFC data: ASCII file in comma separated values format

II.D.2.a.i. Roving diver fish count methods:

The roving diver fish count (RDFC) is a method for counting fish that covers a much larger area than the visual transects and attempts to sample all fish species, instead of just the core set of species counted in visual fish transects. It also has the advantage that multiple counts are done by different observers at a site on most sampling dates. As such, the RDFC data provide more complete information on the total fish community than the fish transect data. Because the RDFC seeks to sample fish within a fixed area, it is tempting to convert these data into densities, but because sampling occurs over a fixed time (30 minutes) there are potential pitfalls with treating these data as density estimates.

The RDFC method produces two indices of fish abundance for each species: an abundance category (single, few, common, many) and a count of the number of individuals observed (see Table 11 for the relationship between categories and counts). The count data is a better index of abundance, but we retain data on abundance category since fish counts were only added to the RDFC protocol in 2003. Prior to that (from 1996 to 2002) only abundance categories were recorded. Even after 2003 there have been a few cases when observers recorded an abundance category instead of a count for a species—usually because there were too many fish for the observer to reliably count all of the species that were present. Fish are identified to species where possible, and in some cases are separated into adults and juveniles or males and females. Although the taxonomic resolution has remained constant over time, in many cases adults and juveniles were counted together prior to 1999 and separated into adult and juvenile categories from 1999 onward.

During the roving diver fish count, divers gradually swim around the transect line, covering the entire transect area in 30 minutes. The sampling area encompasses ten meters on both sides of the 100 m main transect yielding a total sampling area of 2000 m2. Throughout the fish count, each diver attempts to search all habitats (i.e. bottom, midwater, under ledges, canopy, etc.) counting all fish observed. Divers attempted to count each fish only once, although some double counting likely occurs, particularly for more mobile species. Each observer is assigned an individual skill level rating based on their experience, but only data from the most experienced observers (i.e., “expert”) is included here. An “expert” is defined here as an observer who can confidently identify and count all species of fish that commonly occur at the Channel Islands. For each species, the total count or abundance category are recorded by each observer separately—there may be from 1-8 separate expert counts done at a site on one date and from 1-12 expert counts done at each site in a year.

Table 11: Definition of fish abundance categories

Abundance Equivalent Category Count Single 1 Few 2-10

Common 11-100 Many 100+

II.D.2.b. Fish size-frequency data:

Table 12A: Definition of column headers in the fish size-frequency data

Table 12B: Definition of variables in the fish size-frequency data

Fish size-frequency data: ASCII file in comma separated values format

II.D.2.b.i. Fish size-frequency methods:

The KFM collects data on fish size-frequency to help capture changes in the size structure of the fish community that would not be apparent from simple counts. This sampling method is performed during or after the roving diver fish count with a minimum sampling time of 30 minutes. As with the RDFC, the observer will sample as much of the 2000 m2 area (ten meters on either side of the 100 m permanent transect line) as possible. Within this area and time, the sizes of as many individuals as possible are estimated, prioritizing certain species as described below. Observers for this protocol are trained, tested and able to accurately estimate fish sizes underwater to within 20% of the actual total length (TL). Observers estimate the total length of fish to the nearest centimeter searching all habitats (i.e. bottom, midwater, under ledges, water column, canopy, etc.). When schools of fish of the same species are encountered at a site, the mean size of individuals in that group is estimated, as well as the number of fish in the group, and the size of the largest and smallest individuals.

All fish observed are measured except cryptic species (e.g., sculpins and sand dabs) and schooling baitfish (e.g., such as sardines, anchovies and smelt). Some measured species are given priority over others, with species ranked by priority categories from 1 (highest) to 4 (lowest) as indicated in the data set. Fish species are prioritized to ensure that particular target species are sampled even if time does not allow for sampling all fish at the site. As of autumn 2012 it has been possible to sample the fish community completely for the top three priority levels, so the size structure of fish given priority 1, 2 or 3 can be considered complete (i.e., sizes were recorded for all fish of those species encountered).

II.D.2.c. Invertebrate size-frequency data:

Table 13A: Definition of column headers in the invertebrate size-frequency data

Table 13B: Definition of variables in the invertebrate size-frequency data

Invertebrate size-frequency data: ASCII file in comma separated values format

II.D.2.c.i. Invertebrate size-frequency methods: The KFM collects data on the size structure of populations of key invertebrate species to help capture changes in those populations that might not be apparent from simple measures of abundance. The number of individuals measured has varied among species, sites and years. A diver records the sizes of individuals of each target species, completely searching swaths running parallel to the main transect, with the spacing and length of these swaths chosen to gather a representative sample from the entire 100 m transect line. When there are relatively low densities of the target species, the observer conducts one long swath; when densities are very high, observers may sample haphazardly placed areas as small as 0.5 m2. In every case all target organisms

within the chosen area are measured to ensure that an unbiased sample of sizes is obtained. For Stylaster californica, Lophogorgia chilensis, Muricea fruticosa and Muricea californica we report the estimated area of the fan in cm2 based on measurements of height and width (made to the nearest centimeter) and assuming an oval fan shape. All other invertebrate measurements are made in centimeters—measuring the diameter of the test for urchins, the longest dimension of the shell for gastropods and the distance from the mouth to the end of the longest arm for sea stars. Measurements are made to the nearest 0.1 cm in situ with minimal disturbance to the organisms, except for sea urchins, which are lifted (if possible) to search for juveniles under the spine canopy of large adults.

II.D.2.d. Giant kelp size-frequency data:

Table 14A: Definition of column headers in the giant kelp size-frequency data

Table 14B: Definition of variables in the giant kelp size-frequency data

Giant kelp size-frequency data: ASCII file in comma separated values format

II.D.2.d.i. Giant kelp size-frequency methods: The KFM collects data on the size structure of giant kelp, Macrocystis pyrifera, greater than 1 m tall at each site where it is present, to help capture changes in kelp populations that might not be apparent from simple measures of its abundance. Giant kelp sizes are measured within a swath running along the entire length of the permanent 100 m transect. If giant kelp are rare at the site all plants within 10 m of the transect will be sampled, otherwise a narrower swath is sampled, with a width chosen to include at least 100 individuals. Within the sampling area all plants encountered are measured. For each plant, the number of stipes one meter above the bottom is recorded, as well as the greatest diameter of the holdfast, measured to the nearest centimeter. In some cases holdfasts were not measured and the code “NA” indicates this.

II.D.2.e. Giant kelp supplementary density data:

Table 15A: Definition of column headers in the giant kelp supplementary density data

Table 15B: Definition of variables in giant kelp supplementary density data

Giant kelp supplementary density data: ASCII file in comma separated values format

II.D.2.e.i. Giant kelp supplementary density methods: In addition to the data on Macrocystis pyrifera included in the benthic density data set (section II.D.1.a.), the KFM has collected additional data on the abundance of this species since 1996. The benthic density data set, collected since 1982, reports the abundance of giant kelp adults and juveniles, with adults defined as plants with one or more stipes taller than 1 meter. It provides a very useful time series, but because it is derived from the 1 m2 quadrats (section II.D.1.a.ii.) the total area sampled is small (less than 40 m2). Since 1996 the KFM has recorded two additional metrics of kelp abundance, measured in the 5 m2 quadrats (see section II.D.1.a.iii. for sampling methods): the density of adult giant kelp (plants greater than 1 m tall and possessing haptera above the primary dichotomy) and the density of sub-adult giant kelp (plants greater than 1 m tall without haptera above the primary dichotomy). Note that the adult and sub-adult plants counted here would both be considered “adult” in the core benthic density data set. Since 2007 the KFM has also estimated the density of stipes on plants greater than 1 m tall within the 1 m2 quadrats (counting stipes without respect to what plant they are found on. Summer stipe density is a strong predictor of total giant kelp biomass and

27 production (Reed et al. 2009). In the giant kelp supplementary density we report the mean of each measure of density at the site, averaged over all replicates within the site for a given year. Also included are the standard error in density calculated across the sampling replicates, the number of replicates and the area sampled in each replicate.

II.D.2.f. Artificial recruitment module data:

Table 16A: Definition of column headers artificial recruitment module data

Table 16B: Definition of variables in the artificial recruitment module data

Artificial recruitment module data: ASCII file in comma separated values format

II.D.2.f.i. Artificial recruitment module methods: Artificial recruitment modules (ARMs) were placed at 11 sites starting in 1992 and selected organisms accumulating within them have been sampled each summer. The number and location of ARMs varies from site to site (typically 5-15 per site). Each artificial recruitment module consists of a wire cage made of 2” x 4” mesh wire (Figure 7) filled with 20 bricks that serve as recruitment substrates. Bricks are placed in the ARM in five layers with four bricks per layer arranged around the internal perimeter of the ARM, leaving an opening in the center of each row. Bricks are made by cutting a concrete cinder block in half longitudinally to produce 2 bricks, each with a cross section shaped like a lower case "m" (Figure 7). More details on the construction of the ARMs are given in Davis (1995).

To sample each ARM, the bricks are removed from the cage and all target organisms are collected, counted and sized to the nearest mm—measuring the diameter of the test for urchins, the widest dimension of the shell for gastropods and the distance from the mouth to the end of the longest arm for sea stars. Parastichopus parvimensis are not measured but are instead classified as larger or smaller than 10cm. Mean size and the associated standard error are reported for all species except P. parvimensis, for which only the fraction less than and fraction greater than 10cm are reported. If there are 200 or more Strongylocentrotus purpuratus or S. franciscanus per ARM, not all of the ARMs are sampled for these two species. In some cases, organisms are brought to the surface to be measured. After sampling all of the bricks and organisms are replaced in the ARM.

Figure 7: Example of an Artificial Recruitment Module (ARM).

II.D.2.g. Subtidal temperature data:

Table 17A: Definition of column headers subtidal temperature data

Table 17B: Definition of variables in the subtidal temperature data

Subtidal temperature data: ASCII file in comma separated values format

II.D.2.g.i. Subtidal temperature methods: Since 1993 the KFM has used automated loggers to measure water temperature at the monitoring sites over time. At each site, temperature loggers programmed to record temperature are bolted to the substrate. Loggers record temperature between 5 and 24 times a day, depending on logger memory capacity. Records are then averaged over each 24 hour period. The depth of the loggers at each site is listed in Table 1. In most years, two temperature loggers are deployed at each site. Loggers are serviced and data collected once per year during the summer sampling.

Upon retrieval, the data from the two loggers are compared and if the recorded temperatures differ by more than 0.2 C, the loggers are independently checked to determine which is most accurate and data from that logger is used. On the rare occasion that both loggers fail or are missing, no temperature data is entered into the data set for that site and year, so there are a few gaps in the time series. Several brands of temperature logger have been used at different periods, with some overlap, including HoboTemp, StowAway, Tidbit and UTBI Tidbit V2 loggers (all made by the Onset Computer Corporation).

Class III. Data set status and accessibility

A. Latest data update B. Latest metadata update C. Data verification D. Copyright or proprietary restrictions

III.A. Latest data update: The data set may be periodically updated. All updates to the data have been logged in Table 18A. Please check for the latest update before using the data set.

III.B. Latest metadata update: The metadata may also be updated periodically. All updates have been logged in Table 18B. Please check for the latest update before using the data set.

III.C. Data verification: Field sheets were proofed for concerns after every day in the field as well as during data entry.

III.D. Copyright or proprietary restrictions: None

Class IV. Data set structural descriptors

A. Data files B. Metadata tables

IV.A. Data Files:

Benthic density data

Benthic cover data

Fish density data

RDFC data

Fish size-frequency data

Invertebrate size-frequency data

Giant kelp size-frequency data

Giant kelp supplementary density data

Artificial recruitment module data

Subtidal temperature data

IV.B. Metadata Files:

Table 1: Monitoring Sites and Their Coordinates

Table 2: History of special fishing restrictions

Table 3: Species Monitored by Method and Year Monitored

Table 4A: Definition of Column Headers in the Benthic Density Data

Table 4B: Definition of Variables in the Benthic Density Data

Table 7A: Definition of column headers in benthic cover data

Table 7B: Definition of variables in the benthic cover data

Table 9A: Definition of column headers in fish density data

Table 9B: Definition of variables in the fish density data

Table 10A: Definition of column headers in RDFC data

Table 10B: Definition of variables in the RDFC data

Table 12A: Definition of column headers in the fish size-frequency data

Table 12B: Definition of variables in the fish size-frequency data

Table 13A: Definition of column headers in the invertebrate size-frequency data

Table 13B: Definition of variables in the invertebrate size-frequency data

Table 14A: Definition of column headers in the giant kelp size-frequency data

Table 14B: Definition of variables in the giant kelp size-frequency data

Table 15A: Definition of column headers in the giant kelp supplementary density data

Table 15B: Definition of variables in the giant kelp supplementary density data

Table 16A: Definition of column headers in the artificial recruitment module data

Table 16B: Definition of variables in the artificial recruitment module data

Table 17A: Definition of column headers in the subtidal temperature data

Table 17B: Definition of variables in the subtidal temperature data

Table 18A: History of data updates

Table 18B: History of metadata updates

Class V. Supplemental descriptors

A. Location of completed data forms B. Data entry verification procedures C. Publications using the data set

V.A. Location of completed data forms: Original field data forms are archived at the Channel Islands National Park office. All forms are scanned and stored electronically in PDF form on the park’s network server.

V.B. Data entry verification procedures: See III.C.

V.C. Publications using the data set prior to 2013:

Babcock RC, Shears NT, Alcala AC, Barrett NS, Edgar GJ, Lafferty KD, McClanahan TR, Russ GR. 2010. Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proceedings of the National Academy of Sciences 107:18256-61.

Behrens M, Lafferty KD. 2004. Effects of marine reserves and urchin disease on southern Californian rocky reef communities. Marine Ecology Progress Series 279:129-139.

Byrnes J, Stachowicz JJ, Hultgren KM, Randall Hughes A, Olyarnik SV, Thornber CS. 2006. Predator diversity strengthens trophic cascades in kelp forests by modifying herbivore behaviour. Ecology Letters 9:61-71.

Davis G. 1993. Design elements of monitoring programs: the necessary ingredients for success. Environmental Monitoring and Assessment 26:99-105.

Davis GE. 2005. Science and Society: Marine Reserve Design for the California Channel Islands. Conservation Biology 19:1745-1751.

Davis G, Haaker P, Richards DV. 1996a. Status and trends of white abalone at the California Channel Islands. Transactions of the American Fisheries Society 125:42-48.

Davis G, Richards D, Kushner D. 1996b. Kelp forest monitoring design review. Technical Report CHIS- 96-01, Channel Islands National Park, California.

Fancy SG, Gross JE, Carter SL. 2009. Monitoring the condition of natural resources in US national parks. Environmental Monitoring and Assessment 151:161-174.

Gelpi CG, Norris KE. 2008. Seasonal temperature dynamics of the upper ocean in the Southern California Bight. Journal of Geophysical Research 113:1-18.

Graham MH. 2004. Effects of local deforestation on the diversity and structure of southern California giant kelp forest food webs. Ecosystems 7:341-357.

Grantham B, Eckert G. 2003. Dispersal potential of marine invertebrates in diverse habitats. Ecological Applications 13:S108-S116.

Halpern BS, Cottenie K. 2007. Little evidence for climate effects on local-scale structure and dynamics of California kelp forest communities. Global Change Biology 13:236-251.

Halpern BS, Cottenie K, Broitman BR. 2006. Strong top-down control in southern California kelp forest ecosystems. Science 312:1230-2.

Karpov K, Haaker P, Albin D, Taniguchi IK, Kushner D. 1998. The red abalone, Haliotis rufescens, in California: Importance of depth refuge to abalone management. Journal of Shellfish Research 17:863- 870.

Lafferty KD, Kuris AM. 1993. Mass mortality of abalone Haliotis cracherodii on the California Channel Islands: tests of epidemiological hypotheses. Marine Ecology Progress Series 96:239-248.

Lafferty KD, Kushner DJ. 2000. Population regulation of the purple sea urchin, Stronylocentrotus purpuratus, at the California Channel Islands. Proceedings of the Fifth California Islands Symposium:380-381.

Lafferty K. 2004. Fishing for lobsters indirectly increases epidemics in sea urchins. Ecological Applications 14:1566-1573.

Micheli F, Halpern BS. 2005. Low functional redundancy in coastal marine assemblages. Ecology Letters 8:391-400.

Reed DC, Raimondi PT, Carr MH, Goldwasser L. 2000. The role of dispersal and disturbance in determining spatial heterogeneity in sedentary organisms. Ecology 81:2011-2026.

Rogers-Bennett L, Allen B, Davis GE. 2004. Measuring abalone (Haliotis spp.) recruitment in California to examine recruitment overfishing and recovery criteria. Journal of Shellfish Research 23:1201- 1207.

Rogers‐Bennett L, Haaker P, Karpov KA, Kushner DJ. 2002. Using spatially explicit data to evaluate marine protected areas for abalone in southern California. Conservation 16:1308-1317.

Schroeter SC, Reed DC, Kushner DJ, Estes JA, Ono DS. 2001. The use of marine reserves in evaluating the dive fishery for the warty sea cucumber (Parastichopus parvimensis) in California, U.S.A. Canadian Journal of Fisheries and Aquatic Sciences 58:1773-1781.

Schroeter S, Page H, Dugan J, Culver CS, Steele B, Guiterrez R, Richards JB, Kushner D. 2009. Scales of variability in larval settlement within the Channel Islands National Marine Sanctuary and along the mainland coast. Proceedings of the 7th California Islands Symposium:151-160.

Zacharias MA, Kushner DJ. 2006. Sea temperature and wave height as predictors of population size structure and density of Megastraea (Lithopoma) undosa: Implications for fishery management. Bulletin of Marine Science 79:71-82.

Acknowledgements

The design, implementation and data collection of this ecological monitoring program have been almost entirely supported by the United States National Park Service. The California Department of Fish and Wildlife and the U.S. Department of Commerce’s National Oceanographic and Atmospheric Administration have both cooperated greatly with this program. As of 2011, 377 scientists and resource managers have collected data underwater as part of this program. There have also been a large number of other scientists that helped to develop and improve the experimental design and resulting monitoring protocols. In addition to people listed as divers, other support staff including boat captains, dive officers and other topside personnel have all contributed to the success and longevity of the Kelp Forest Monitoring Program since 1982. Of particular importance was the founder of the monitoring program Gary E. Davis who had the vision to develop and support a program that continues to be a valuable component to resources management and support our long-term understanding of kelp forest ecology.

During the preparation of this data paper A.R. and J.M. were supported with funds from the U.S. Department of the Interior, Bureau of Ocean Energy Management, Environmental Studies Program, Washington, DC under Cooperative Agreement Number M11AC00012. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the opinions or policies of the U.S. Government. Mention of trade names or commercial products does not constitute their endorsement by the U.S. Government. This product (article, paper, etc.) has been peer reviewed and approved for publication consistent with USGS Fundamental Science Practices (http://pubs.usgs.gov/circ/1367/).

Literature cited

Airamé S, Dugan J, Lafferty K, Leslie H, McArdele D, Warner R. 2003. Applying ecological criteria to marine reserve design: a case study from the California Channel Islands. Ecological Applications 13:170–184.

Cavanaugh KC, Siegel DA, Reed DC, Dennison PE. 2011. Environmental controls of giant kelp biomass in the Santa Barbara Channel, California. Marine Ecology Progress Series 429:1–17.

Davis GE. 1995. Recruitment of juvenile abalone (Haliotis spp.) measured in artificial habitats. Marine and Freshwater Research 46:549–454.

Davis GE. 2005. Science and society: marine reserve design for the California Channel Islands. Conservation Biology 19:1745–1751.

Davis G, Richards D, Kushner D. 1996. Kelp forest monitoring design review. Park Service, Channel Islands National Park. Channel Islands National Park, California. 24 pp.

Engle JM. 1994. Perspectives on the structure and dynamics of nearshore marine assemblages of the California Channel Islands. Proceedings of the Fourth California Islands Symposium:13-26.

Lafferty K. 2004. Fishing for lobsters indirectly increases epidemics in sea urchins. Ecological Applications 14:1566–1573.

Reed D, Raimondi P, Carr M, Goldwasser L. 2000. The role of dispersal and disturbance in determining spatial heterogeneity in sedentary organisms. Ecology 81:2011–2026.

Reed DC, Rassweiler A, Arkema KK. 2009. Density derived estimates of standing crop and net primary production in the giant kelp Macrocystis pyrifera. Marine Biology:2077–2083.

Appendix B

All files and folders referenced in Appendix B are found in the file: OCS Study BOEM 2019-063 Appendix B files.zip

A multi-decade time series of kelp forest community structure at San Nicolas Island, California

Michael C. Kenner, James A. Estes, M. Tim Tinker, James L. Bodkin, Robert K. Cowen, Christopher Harrold, Brian B. Hatfield, Mark Novak, Andrew Rassweiler, Daniel C. Reed

2013. Ecology 94: 2654

INTRODUCTION

Reefs at San Nicolas Island are an ideal location for studying the dynamics of shallow subtidal communities. San Nicolas is the most remote of California’s Channel Islands, and its relative isolation somewhat protects it from the terrestrial runoff and recreational use faced by other reefs in Southern California. Although the island is small, its gently sloping subtidal shelf provides substantial shallow area for kelp forests and other rocky reef communities. Further, reefs at San Nicolas Island have been the focus of more than three decades of ecological monitoring, and the resulting data represent a valuable tool for understanding kelp forest community dynamics.

The data presented here come from a monitoring program that was initiated in 1980 in anticipation of the potential reintroduction of sea otters (Rathbun et al. 1990, Rathbun and Benz 1991). San Nicolas Island had been identified as a potential site for the establishment of a new breeding population of sea otters (sea otters from central California were translocated to the island starting in 1987). Subtidal monitoring was initiated in autumn of 1980 at six sampling stations positioned around the island to capture communities exposed to a range of wave exposures and other oceanographic conditions (Table 1). One additional station was added in 1986 (chosen to add a location where the bottom had very little physical relief). The intent of this monitoring was to provide baseline information on ecological community structure in order to better evaluate the effects of sea otter introduction. Each station has been sampled twice a year (with a few exceptions). This period of monitoring covers substantial variability in environmental conditions, including two of the strongest El Nino’s on record, years with high and low waves, and major shifts in the Pacific decadal oscillation.

In this paper, we present time-series data on the abundance of more than 200 species of fish, algae and invertebrates. These data were collected at each permanent sampling station using a combination of swath counts and point-contact sampling. In addition, we present measurements of the sizes of Macrocystis pyrifera (giant kelp), a key foundation species. The long time series and biannual sampling make these data appropriate for exploring ecological dynamics on multiple time scales. The high taxonomic

resolution of the sampling provides a detailed picture of community structure over time, and the fact that fixed quadrats and swaths have been sampled repeatedly means that changes in the community can be examined without the obscuring effects of sampling error. Taken together, these characteristics make these data valuable for investigating the ecology of giant kelp forests, particularly with respect to patterns and drivers of community dynamics across a wide range of physical and biological conditions

METADATA CONTENTS

Class I. Data set descriptors E. Data set identity F. Data set identification code G. Data set description 1. Data set originator 2. Abstract H. Key words

Class II. Research origin descriptors E. Project description 1. Objectives 2. System description 3. Guidelines for use of data F. Monitoring design 6. Site selection and layout 7. Human impacts 8. Monitoring method selection 9. Choice of species to be monitored 10. Monitoring design metadata G. Data processing and presentation 4. Core community structure data sets 5. Supplementary data sets 6. Additional documentation and data H. Monitoring methods and data 1. Core community structure data sets: d. Benthic density data i. Benthic density summary procedures ii. Swath count methods e. Benthic cover data i. Benthic cover summary procedures ii. Point contact monitoring methods f. Fish density data i. Fish density summary procedures ii. Fish transect monitoring methods 2. Supplementary data sets a. Giant kelp size-frequency data i. Giant kelp size-frequency field methods

Class III. Data set status and accessibility E. Latest data update F. Latest metadata update G. Data verification H. Copyright or proprietary restrictions

Class IV. Data set structural descriptors A. Data files B. Metadata tables

Class V. Supplemental descriptors D. Location of completed data forms E. Data entry verification procedures F. Publications using the data set

Class I. Data set descriptors

A. Data set identity B. Data set identification code C. Data set description 1. Data set originator 2. Abstract D. Key words

I.A. Data set identity: Rocky reef community structure at San Nicolas Island, California

I.B. Data Set Identification Code: KennerEtAl2013-SNI

I.C. Data set description

I.C.1. Data set originators

Individual: James A. Estes Role: Initial principal investigator Organization: University of California, Santa Cruz Position: Professor of Ecology and Evolutionary Biology Address: Long Marine Lab, 100 Shaffer Road, Santa Cruz, CA 95060 Phone: 831-459-2820 Email: [email protected]

Individual: Michael C. Kenner Role: Chief field biologist Organization: University of California Santa Cruz Position: Biologist Address: Long Marine Lab, 100 Shaffer Road, Santa Cruz, CA, 95060 Phone: 831-254-5184 Email: [email protected]

Individual: M. Tim Tinker Role: Current principal investigator Organization: US Geological Survey, Western Ecological Research Center Position: Research Biologist Address: Long Marine Lab, 100 Shaffer Road, Santa Cruz, CA, 95060 Phone: 831-459-2357 Email: [email protected]

I.C.2. Abstract:

San Nicolas Island is surrounded by broad areas of shallow subtidal habitat, characterized by dynamic kelp forest communities that undergo dramatic and abrupt shifts in community composition. Although these reefs are fished, the physical isolation of the island means that they receive less impact from human activities than most reefs in Southern California, making San Nicolas an ideal place to evaluate alternative theories about the dynamics of these communities. Here we present monitoring data from seven sampling stations surrounding the island, including data on fish, invertebrate and algal abundance. These data are unusual among subtidal monitoring data sets in that they combine relatively frequent sampling (twice per year) with an exceptionally long time series (since 1980). Other outstanding qualities of the data set are the high taxonomic resolution captured and the monitoring of permanent quadrats and swaths where the history of the community structure at specific locations has been recorded through time. Finally, the data span a period that includes two of the strongest ENSO events on record, a major shift in the Pacific decadal oscillation and the reintroduction of sea otters to the island in 1987 after at least 150 years of absence. These events provide opportunities to evaluate the effects of bottom up forcing, top down control and physical disturbance on shallow rocky reef communities.

I.E. Key words: long-term monitoring, California Channel Islands, rocky reef ecology, sea otters

Class II. Research origin descriptors

A. Project description 1. Objectives 2. System description 3. Guidelines for use of data II.A.1. Objectives: In the early 1980s, as part of the United States Fish and Wildlife Service’s research program on southern sea otters, it was decided that a breeding population of sea otters should be established at San Nicolas Island to reduce the species’ vulnerability to extinction caused by increasing threats from human activities. San Nicolas was selected because of its isolation and high availability of suitable habitat. In anticipation of the establishment of a new population, the U.S. Fish and Wildlife Service and the University of California, Santa Cruz established six permanent subtidal sampling stations around San Nicolas Island in 1980. One additional station was set up in 1986. It was chosen because the site had virtually no bottom relief and thus kelp, when present, would represent the only three dimensional structure affecting fish density. The study was designed to examine the possible impact of a translocated sea otter population on the local reef communities. The project is currently conducted by the U.S. Geological Survey - Western Ecological Research Center, U.C. Santa Cruz Field Station.

II.A.2. System description: San Nicolas Island (Ventura County, CA, USA) is the most remote of southern California’s Channel Islands, both from shore and from the nearest neighboring island, lying approximately 140 km west-south-west of Los Angeles (~110 km offshore, 33°15' N, 119°30' W). San Nicolas is influenced both by the cold southward flowing California Current and the warmer northward edge of the California counter-current; consequently, its kelp forest communities contain biota characteristic of both more northern and more southern regions. The island itself is relatively small (less than 60 km2 in area) but because of its gently sloping shelf of low-relief soft-sedimentary rock, it has the greatest area of kelp beds of any of the Channel Islands (30% of the total kelp coverage throughout the islands; Engle 1994). San Nicolas is located far from the wave shadows of other islands and the shape of its coastline provides little protection from the swell, so most of its shallow-water habitats are exposed to wave disturbance. Because of its isolation from the populated mainland, shallow subtidal communities at San Nicolas are less exposed to human influences than communities on many reefs in Southern California. However, reefs surrounding the island are the focus of several active commercial fisheries (particularly for sea urchins, lobster and rockfish), modest recreational fishing pressure (particularly for rockfish, ocean whitefish and sheephead) and are potentially impacted by the US Navy’s operations on the island.

II.A.3. Guidelines for use of data: This data set is part of an ongoing monitoring program at San Nicolas Island run by the United States Geological Survey Western Ecological Research Center (USGS-WERC).

B. Monitoring Design 1. Site selection and layout 2. Human impacts 3. Monitoring method selection 4. Choice of species to be monitored 5. Monitoring design metadata II.B.1. Site Selection and Layout: Seven stations have been chosen for monitoring (six in 1980 and one in 1986). These stations are placed around San Nicolas to provide data from each side of the island and to be representative of the available kelp forested habitat. A map of the stations is given in Figure 1. Three

stations are isolated (Sandy Cove, Nav Fac and Daytona), while the other four were placed as two pairs. The pair of stations at West End was chosen to represent an urchin dominated area of reef and a kelp dominated area (although the communities at these stations have changed through time). These two stations are contiguous (the 50 m end of West End Kelp is the zero bolt of West End Urchins). The pair of sampling stations at Dutch Harbor was chosen as experimental (west) and control (east) stations for a sheephead removal experiment run in 1980 through 1982 (Cowen 1983). These two stations are separated by ~140 m. The latitude and longitude of each sampling station are given in Table 1. With a few exceptions, each station has been monitored twice a year (in spring and autumn) since its establishment.

Figure 1: Map of SNI monitoring stations.

At each station, monitoring is focused on a permanent 50-meter transect marked by stainless steel eyebolts placed at 5 m intervals. Each transect is placed approximately perpendicular to the coastline between 10 and 15 meters in depth. Associated with each transect are five permanent 10 m by 2 m swaths running perpendicular to it at fixed intervals. The original orientation of swaths (to the left or right of the transect) was chosen at random. Additionally, ten 1 m2 quadrats are permanently placed one meter to the right and left of the main transect along its length. These quadrats mark locations where biota is sampled using fixed point contact methods. Finally, five 50 m fish transects are associated with each sampling station. Each fish transect starts at one of the permanent eyebolts and is laid out along a specific compass heading. Figure 2 illustrates the layout of quadrats and swaths along a typical transect.

Figure 2. An example of the layout of quadrats and swaths at a typical sampling station.

II.B.2. Human impacts: A number of significant anthropogenic impacts have occurred at the monitoring stations since their establishment. Perhaps the most important was the translocation of southern sea otters (Enhydra lutris nereis) from central California to SNI in 1987-1990. During this period, 140 sea otters were released at SNI. Although there was relatively low retention of these translocated animals, a population of ~15 adult animals remained at the island from 1990 to 1998, and their numbers have been increasing consistently since (Hatfield 2003)

Sea otters are not evenly distributed around the island. Sea otter surveys have been done three to five times per year since 1995, and the sampling date with the highest number of sea otters observed is a good measure of their abundance in space and time (Rathbun et al. 2000). When the coastline is divided into three roughly equivalent segments: the west (West Wreck to Thousand Springs, ~ 11.5 km), the north (Thousand Springs to Sand Spit, ~12 km) and the south (Sand Spit to West Wreck, ~ 13 km), it is apparent that independent sea otters (e.g. older juveniles and adults) are consistently abundant in the west and rare in the north (Table 2). This pattern is also observed in sea otter pups, which are common in the west, occasional in the south and effectively absent from the north (Table 3).

Several other human impacts at individual sampling stations are also worth noting. In September 1980, as part of an experimental manipulation, 220 sheephead (Semicossyphus pulcher) were removed from Dutch Harbor West. Cowen (1983) estimated that this manipulation eliminated all sheephead at the site. This manipulation was maintained for 2 years, with additional sheephead removed at bimonthly intervals (10- 20 removed per visit). In addition to this scientifically motivated removal, substantial recreational and

commercial fishing occurs at the island. Although fine scale spatial data on fishing pressure are not available, several long term data sets document changes in fishing in southern California (Perry et al 2010).

The U.S. Navy’s operations on San Nicolas also had significant effects on nearshore kelp forests. Until 2005 when the new pier opened, the navy brought barges in on the beach at Daytona Beach resulting in considerable sand movement, and relatively heavy sediment loads at that site. Since 2005, this practice has stopped.

In addition to human impacts, other factors have affected the continuity of data collection. In particular, the transect at Daytona could not be located in the fall of 1983, probably because of the sand input described above. Swaths 39R and 45R were never relocated and were eventually replaced with 22L and 39L in fall of 1985.

II.B.3. Monitoring method selection: The great diversity of organisms and physical habitats associated with kelp forests requires multiple sampling approaches to effectively monitor species population dynamics. Sampling methods were designed to maximize accuracy and precision while preserving an observer’s ability to efficiently sample many target species at once. Counts within fixed areas are made to describe the abundance of fish as well as large invertebrates and algae. Estimates of percent cover are used to measure the abundance of species for which individuals are numerous, small or otherwise difficult to count.

II.B.4. Choice of species to be monitored: The monitoring captures the abundance of most algae, invertebrates and fish larger than a few centimeters, including all primary benthic spaceholders. The species sampled and data sets in which they appear are found in Table 4.

II.B.5. Monitoring design metadata: The following metadata are descriptive of the monitoring methods and context and are available below as comma separated values files.

Table 1: Monitoring stations and their Coordinates

Table 2: Abundance of independent sea otters

Table 3: Abundance of sea otter pups

Table 4: Species sampled

C. Data processing and presentation 1. Core community structure data sets 2. Supplementary data sets 3. Additional documentation and data II.C.1. Core community structure data sets: The subtidal monitoring data presented here describe the algal, invertebrate and fish components of the kelp forest community at as high a taxonomic resolution as is practical. We present this information in four “core community structure data sets”: benthic density (calculated from counts of organisms within fixed 10 m x 2 m swaths), benthic cover (calculated from point contact data in fixed 1 m X 1 m quadrats), midwater fish density (calculated from 50 m x 5 m fish transects), and benthic fish density (calculated from 50 m x 2 m fish transects). Each of these data sets is presented in both summarized form (with means and standard errors for each station) and in raw form (with counts from each sampling replicate reported separately).

II.C.2. Supplementary data sets: In addition to the core data discussed above, supporting data on Macrosystis pyrifera populations are also included. The basal diameters of a subset of giant kelp plants are measured and, in some cases, stipes have also been counted. Additionally, some giant kelp plants have been given unique tags, allowing the size of individual plants to be tracked over time. These data are collected biannually with the core community data.

II.C.3. Additional documentation and data: The subtidal data presented here are part of a larger monitoring program that includes intertidal community data collected twice a year and sea otter population surveys done three to five times a year. Users interested in these data should contact:

M. Tim Tinker USGS-WERC, Santa Cruz Field Station, Long Marine Lab 100 Shaffer Road, Santa Cruz, CA, 95076 [email protected]

D. Monitoring methods and data 1. Core community structure data sets: a. Benthic density data i. Benthic density summary procedures ii. Swath count methods b. Benthic cover data i. Benthic cover summary procedures ii. Point contact monitoring methods c. Fish density data i. Fish density summary procedures ii. Fish transect monitoring methods

II.D.1.a. Benthic density data

Table 5A: Definition of column headers in the benthic density raw data

Table 5B: Definition of variables in the benthic density raw data

Benthic density raw data: ASCII file in comma separated values format

Table 6A: Definition of column headers in the benthic density summary data

Table 6B: Definition of variables in the benthic density summary data

Benthic density summary data: ASCII file in comma separated values format

II.D.1.a.i. Benthic density summary procedures: Density data for solitary benthic algae and macroinvertebrates are collected by counting the number of individuals in a fixed swath and dividing that count by the area sampled. For each sample date and station, density estimates are reported for each swath in the “raw” data set and as a mean and standard error across all swaths in the “summary” data set.

II.D.1.a.ii. Swath count methods: 19 target species (Table 4) are sampled by counting the number of individuals of each that occur within five 10 m by 2 m swaths at each sampling station. Swaths are laid out perpendicular to the main transect at specified intervals (see Figure 2 for an example of swaths

placement). On each swath, divers count seven species or categories of brown algae (primarily kelps) and 12 species of macro-invertebrates. Only organisms that can be seen without extensive or destructive searching are counted.

II.D.1.b. Benthic cover data

Table 7A: Definition of column headers in the benthic cover raw data

Table 7B: Definition of variables in the benthic cover raw data

Benthic cover raw data: ASCII file in comma separated values format

Table 8A: Definition of column headers in the benthic cover summary data

Table 8B: Definition of variables in the benthic cover summary data

Benthic cover summary data: ASCII file in comma separated values format

II.D.1.b.i. Benthic cover summary procedures: Percent cover data for the benthic community of algae and sessile invertebrates are collected using a fixed point contact method in which a series of points is superimposed over the bottom, and the species intersecting each point is recorded. The number of times each species is contacted is divided by the total number of points sampled and multiplied by 100 to give the percent of the bottom occupied by that species. Percent cover of each species is calculated separately for each of the 10 1 m2 quadrats. For each sample date and station, percent cover estimates for each taxon are reported for each quadrat in the “raw” data set and as a mean and standard error calculated across all quadrats in the “summary” data set. Note that the raw (quadrat level) estimates of cover are based on 20 points each, so give cover estimates in increments of 5% (one point => 5% cover). The summary (station level) data gives cover estimates in increments of .05% (one point => .05% cover). Thus, while the raw data are very useful for looking at small scale associations between common species, the increased precision given by the summary data make them more appropriate for describing population dynamics, particularly of rare species.

II.D.1.b.ii. Point contact monitoring methods: Benthic percent cover data are collected in 1 m2 permanent quadrats, with a quadrat placed 1 m to the left or right of the main 50 m transect at 10 fixed locations. Within each quadrat 20 points are distributed in a fixed pattern. The diver imagines a line running vertically through each point up to one meter above the substratum and identifies all organisms that intersect this imaginary line. The diver counts each species only once per point even if multiple individuals of the same species intersect that point. Because of this method of counting multiple layers, total cover of all species can exceed 100% (and usually does), but the cover of any individual species cannot be > 100%. The list of taxa recorded is open and ranges from actual individual species to species groups like “orange encrusting sponge”. A few organisms which could not be identified have been recorded as “Unidentified spp.” Substrate type is also recorded if exposed.

II.D.1.c. Fish density data

Table 9A: Definition of column headers in the midwater fish density raw data

Table 9B: Definition of variables in the midwater fish density raw data

Midwater fish density raw data: ASCII file in comma separated values format

Table 10A: Definition of column headers in the midwater fish density summary data

Table 10B: Definition of variables in the midwater fish density summary data

Midwater fish density summary data: ASCII file in comma separated values format

Table 11A: Definition of column headers in the benthic fish density raw data

Table 11B: Definition of variables in the benthic fish density raw data

Benthic fish density raw data: ASCII file in comma separated values format

Table 12A: Definition of column headers in the benthic fish density summary data

Table 12B: Definition of variables in the benthic fish density summary data

Benthic fish density summary data: ASCII file in comma separated values format

II.D.1.c.i. Fish density summary procedures: Fish density data have been collected since autumn 1981 by counting the number of individuals within fixed transects and dividing that count by the area of bottom sampled. At each sampling date, fish are counted in five fish transects per sampling station. A diver passes over each transect twice, counting midwater fish in a 5 m wide transect on the first pass and benthic fish in a 2 m wide transect on the second. These data are presented separately here, because the same individuals could be present in both counts depending on their behavior. Counts are presented as densities per m2 of horizontal area sampled. For each sample date and sampling station, density estimates are reported for each fish transect separately in the “raw” data sets and as a mean and standard error across all transects in the “summary” data sets. On one occasion (at Daytona Beach in October 1988) a huge school of California anchovy (Engraulis mordax) was observed on two transects, but they were too numerous to get an estimate of their abundance. They have been omitted from the data.

II.D.1.c.ii. Fish transect monitoring methods: Five 50 m visual fish transects are sampled at each station on each sampling date. Each of these transects is defined by a beginning point on the main transect and a compass heading, so that the same areas are counted each sampling period (see Table 11 for fish transect placement). On each transect, two counts are made. First the observer swims the entire transect about 2-3 m off the bottom, sampling a 5 m wide transect and counting all fish found between the surface and 2 m above the bottom. Sheephead and all schooling fish (e.g. blacksmith) are counted even if they are within 2 m of the bottom. Fish observed in this count are recorded as midwater fish; the total horizontal area sampled is 250 m2. After reaching the end of the transect, the observer works backwards more slowly, counting all fish found in a 2 m wide swath within 2 m of the bottom. Midwater fish (such as sheephead) are excluded from this count, even if they are found within 2 m of the bottom. This is slow and careful work, because many of these fish are mixed in with the benthic algae, so despite the smaller area sampled, this second count typically takes longer than the first. Fish observed in the second count are recorded as benthic fish. The total horizontal area sampled is 100 m2. In each count, an open species list is used, so all fish encountered are sampled. Fish are identified to the highest taxonomic resolution possible in the field, typically to species or genus. In both counts, juveniles are scored separately when observed, and male and female sheephead (Semicossyphus pulcher) are recorded separately.

Table 13: Starting point on the main transect and orientation of fish transects

Station StationName 1 2 3 4 5

1 Nav Fac 0 m E 10 m W 20 m E 30 m W 40 m W

2 West End, Urchin 0 m N 10 m N 20 m N 30 m N ON

3 West End, Kelp 0 m S 10 m N 20 m S 30 m N ON

4 West Dutch Harbor 0 m W 10 m W 30 m W 45 m W ON

5 East Dutch Harbor 0 m E 10 m W 20 m E 30 m W 40 m E

6 Daytona Beach 0 m E 0 m W 10 m E 20 m W 30 m E

7 Sandy Cove 0 m W 10 m W 20 m E 30 m E ON

Note: “ON” means the fish transect was performed along the main transect, N, E, S and W refer to compass headings of 0, 90, 180 and 270 degrees respectively.

II.D.2. Supplementary data sets a. Giant kelp size-frequency data i. Giant kelp size-frequency field methods

II.D.2.a. Giant kelp size-frequency data:

Table 14A: Definition of column headers in the giant kelp size-frequency data

Table 14B: Definition of variables in the giant kelp size-frequency data

Giant kelp size-frequency data: ASCII file in comma separated values format

IV.D.2.a.i. Giant kelp size-frequency methods: Beginning in spring 1981, the basal diameters (holdfast width) of all Macrocystis pyrifera plants greater than 1 m tall encountered on swaths have been measured. In some cases the number of stipes on each plant has been counted as well. Stipe counts are typically done on the first two swaths at each sampling station with kelp present, but counts have not been made in all such cases. To obtain information on patterns of recruitment, mortality and lifespan of giant kelp, some individuals were tagged with unique numbers, written on plastic tape and cable-tied loosely around the primary dichotomous branch above the holdfast and followed over time. The number of tagged plants varied through time, sometimes including all plants > 1 m tall, and at other times including only the first 10 plants encountered on each swath. Tagging was discontinued in 2004 to reduce overall sampling effort. Missing holdfast sizes, frond counts or tag numbers are noted with “NA” in the data set.

Class III. Data set status and accessibility

I. Latest data update J. Latest metadata update K. Data verification L. Copyright or proprietary restrictions

III.A. Latest data update: The data set may be periodically updated. All updates to the data have been logged in Table 15A. Please check for the latest update before using the data set.

III.B. Latest metadata update: The metadata may also be updated periodically. All updates have been logged in Table 15B. Please check for the latest update before using the data set.

III.C. Data verification: Field data sheets are proofed for accuracy after every day in the field as well as during data entry.

III.D. Copyright or proprietary restrictions: None

Class IV. Data set structural descriptors

C. Data files D. Metadata tables

IV.A. Data Files:

Benthic density raw data

Benthic density summary data

Benthic cover raw data

Benthic cover summary data

Midwater fish density raw data

Midwater fish density summary data

Benthic fish density raw data

Benthic fish density summary data

Giant kelp size-frequency data

IV.B. Metadata Files:

Table 1: Monitoring Stations and Their Coordinates